AI community pushes back on personal attacks against evals lab METR

austinc3301 · x · 2026-09-16

Practitioners are speaking out against recent attacks on AI safety evals nonprofit METR. While some substantive criticism exists, most of it targets the appearance of METR staff rather than their work. One commenter calls METR the most competent team in the world at what they do, warning that we shouldn't burn the scarce expertise we have in a field that needs far more organizations like it.

Original post →

More from Safety

Safety channel →